|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
66.68.73.9
I'm building an integrated transformer coupled DHT differential amp, and am interested in exploring the WE harmonic equalizer circuit. I'd like to explore the measured effects of different cap and resistor values with this circuit, on 3rd order distortion, in the individual stages of my amp. Would WinAudioMLS be an effective tool to measure the effects, and if not what would? BTW if a get any results I'll post.
Follow Ups:
John Atwood (over on the ClariSonus blog) has done some preliminary research, and found a mild degree of load sensitivity for the Harmonic Equalizer (or Harmonic Balancer) circuit - there's also some tuning required for the distortion profile of the tube. The tuning wasn't very critical, though - the lowest distortion tubes (DHT's) were at one end of the adjustment, and triode-connected pentodes and IDHT's were towards the middle.The circuit does strongly reduce not only 3rd-harmonic, but all odd harmonics, out to the 9th and greater. Interestingly, in an obscure publication in the 1935 American Projectionist magazine, Bell Labs/WE mentions that feedback is most suitable for high-distortion amplifiers - referring to their own 91A SET amplifier, which does use feedback - but the Harmonic Equalizer was reserved for lower-distortion amplifiers, such as their 86 and 92 amplifiers. This is the opposite of the conventional wisdom we have today of non-feedback SET's and feedback PP amplifiers.
John and Mark Kelly came up with an interesting model, equating the balanced amplifier to a balanced modulator, with the WE circuit controlling the common-mode return current between plate and cathode. John also found the relevant patent: 1,970,325, filed by by John G. Kreer, assigned to Bell Labs, and granted in August 14, 1934.
If I wasn't working on getting ready for the RMAF, I'd have posted this already on the blog.
.
Lynn- You should know that I have very little formal training in electronics (just what I picked up in the IBEW apprenticeship program). I'm trying to bootstrap myself to some limited understanding. It's been a long and slow process!:-) I know that the WEHE circuit nulls other odd ordered harmonics, I just assumed that what nulled 3rd, would probably null other odd orders as well. I could be wrong about that, you know what they say about that word assume.I would like some guidance as to what equipment would be necessary to measure the results. I know a spectrum analyzer would be ideal, but can't afford one of those puppies. The test equipment I do have, and am just beginning to learn about: 60MHz scope, HP 333/A distortion analyzer, and a Heathkit IG-18 signal generator, plus various hand held meters.
I saw this WinAudioMLS software (I'll post link), and thought with a decent exterior sound card, that this might do what I want. It'd also be a cool tool for speaker measurement/crossover tweaks, as well as room measurement, which would be a nice plus. However, if I can measure odd order harmonic distortion with what I already have (I doubt it's accurate enough), I'd like to save the bucks. Also, I don't know if this software will do the trick, and don't know what kind of I/O devices it would take to interface an amp with the computer. I was hoping someone on this forum had experience with the software, but so far no luck.
I've been following your Amity-Aurora projects over the years, and really admire what you're doing, and especially your spirit to share this valuable information. It's inspired me to take on this project. I've been scheming on this design for over a year now, and have finally decided (after many changes) on this topology: input/phase splitter transformer > secondary to ladder stereo attenuator > 5687 Dif. preamp stage > IT > 801A Dif driver stage > IT > Medium voltage (~ 800V anode, ~ 95V cathode) 845 Dif. power stage. The (6) separate power supplies for each stage and channel will be more conventional (relative low DCR CLC) than yours. The regulated CCS power supplies are too difficult for me to implement, but I'm thinking CCSs on the cathodes would be interesting. Planning on 4 chassis, left signal, right signal, left power supply, right power supply. I have about 90% of the necessary parts, and will start bread boarding the preamp stage soon.
I'd like to explore the effects of the WEHE on all three stages, as I don't really know, and can't seem to find, what cap/resistor values would be optimal with these tubes. So why not take a bunch of caps, and a pot and find out? Thanks again for all your work, knowledge, and willingness to share.
twystd
As you can imagine, there's a spectrum of results that you get, depending on the signal fed to the common cathodes. On one end, there's a current source, which forces the circuit into full differential, and prevents any kind of Class AB mode of operation. On the other end, the circuit is running in parallel, with generous bypassing either to ground or to the common B+ point on the center-tap of the output transformer.Since the error term we're trying to minimize is 3rd harmonic, and all odd harmonics, pure differential operation doesn't necessarily have the lowest distortion. John Atwood found to his surprise that an unbypassed current source actually had the greatest 3rd-harmonic distortion, in addition to the lowest power and sharpest overload characteristic.
In terms of reduction of 2nd harmonic, all of the balanced circuits were fairly similar - but it doesn't matter that much, since 2nd harmonic is much less audible than 3rd harmonic. Where the circuits differed was the proportion of 3rd and other other higher harmonics.
John also found that tubes with ultra-low distortion, or more accurately, tubes with the lowest amount of high-order terms, had the lowest 3rd-harmonic with the pot set all the way over to complete bypassing. Tubes with a bit more high-order content had the lowest amount of 3rd-harmonic had an optimum adjustment with a 1~3 K of resistance in the circuit.
The overall results of the test indicated the phase and amplitude of the common-mode term on the common cathode are what controls the 3rd-harmonic. With an unbypassed current source, the common term had one polarity, with the cathodes connected to the B+ terminal, the polarity was reversed, and with a judiciously chosen resistor, the term can be reduced to zero. Again, this is the COMMON-MODE term, which is all 3rd-harmonic and higher in character.
(By the way, this is why it is fallacious to think that a Class A PP power supply sees no deviation in current demand. This would only be true if the tubes were distortionless, which obviously isn't true. The distortion in the tubes appears as 3rd-harmonic on the power supply, and also appears in the common-cathode circuit.)
This series of common-mode tests, by the way, was only in private correspondence between me, John Atwood, and Mark Kelly, so you won't see it in the Clarisonus blog right away (maybe after the RMAF when we all have a bit more time).
As for software, any good FFT program will show you what you need. 48 kHz and 16 bits is plenty for measuring tube amps. Be sure to protect your sound card with back-to-back protection diodes, and possibly transformer coupling to chase out any ground loops, which can be hard to remove otherwise.
.
This post is made possible by the generous support of people like you and our sponsors: